Table 2.

Performance of machine learning in predicting hourly PM₁₀.

Time	Random Forest		RNN
	R	RMSE	R	RMSE
Beijing, BTH
2015.1	0.85	56.3	0.86	58.6
2016.1	0.91	60.2	0.92	41.0
2017.1	0.86	77.4	0.88	82.9
2018.1	0.74	34.5	0.75	34.8
2019.1	0.85	35.9	0.85	35.8
Tianjin, BTH
2015.1	0.87	57.9	0.92	51.2
2016.1	0.89	50.5	0.89	50.3
2017.1	0.88	51.9	0.88	51.9
2018.1	0.78	26.4	0.82	23.5
2019.1	0.79	36.8	0.81	42.9
Shanghai, YRD
2015.1	0.86	42.7	0.91	44.2
2016.1	0.83	30.3	0.88	27.4
2017.1	0.79	25.6	0.79	22.6
2018.1	0.86	27.8	0.88	28.7
2019.1	0.85	34.0	0.87	31.9
Nanjing, YRD
2015.1	0.81	59.9	0.84	69.9
2016.1	0.69	48.4	0.84	70.2
2017.1	0.81	32.6	0.97	29.1
2018.1	0.80	53.2	0.81	55.6
2019.1	0.83	44.5	0.83	50.1
Hangzhou, YRD
2015.1	0.89	39.2	0.90	39.7
2016.1	0.84	34.8	0.87	30.9
2017.1	0.80	34.2	0.84	32.4
2018.1	0.82	30.0	0.87	28.3
2019.1	0.61	45.9	0.61	49.7
Guangzhou, PRD
2015.1	0.86	23.5	0.930	16.2
2016.1	0.86	17.5	0.882	16.4
2017.1	0.88	27.9	0.895	29.9
2018.1	0.90	33.5	0.921	32.9
2019.1	0.89	27.9	0.87	38.4
Shenzhen, PRD
2015.1	0.80	23.7	0.79	28.3
2016.1	0.73	16.4	0.75	16.0
2017.1	0.79	12.2	0.80	13.8
2018.1	0.89	17.7	0.90	18.3
2019.1	0.81	25.3	0.81	27.7
Chengdu, SCB
2015.1	0.84	78.1	0.86	77.4
2016.1	0.84	31.5	0.79	36.7
2017.1	0.70	96.7	0.69	107.7
2018.1	0.78	37.6	0.86	31.3
2019.1	0.57	34.9	0.68	36.8
Chongqing, SCB
2015.1	0.82	69.7	0.81	75.1
2016.1	0.60	37.2	0.63	34.8
2017.1	0.79	41.2	0.85	38.6
2018.1	0.80	35.3	0.89	26.6
2019.1	0.64	49.2	0.69	48.6